UL AI Certificate: What I Learned and How It Changed My Builds

I spent Sep 2025 – Jan 2026 completing an AI certificate programme at the University of Limerick.

I decided not to continue to the full MSc programme afterwards — not because AI is less important, but because I wanted to bias toward building real systems while learning in public.

What I wanted from structured study

After years in cloud support and DevOps, I noticed a pattern:

• I’m good at debugging systems.
• I’m good at shipping products.
• But I wanted more structured time for AI fundamentals so I could design better LLM systems beyond “prompting”.

The certificate gave me the discipline to revisit core ideas and connect them directly to practical projects.

How it shows up in my projects

1) RAG Troubleshooting Assistant

The certificate pushed me to think more carefully about:

• evaluation (not just “looks good”)
• retrieval quality (chunking, embeddings, relevance)
• reliability and UX for technical decision-making

2) VideoTranslate

A classic “cloud + AI glue” project:

• event-driven architecture
• automation-first design
• outputs that can be reused downstream (search, subtitles, knowledge bases)

3) ArgumentLab

The most personal project:

• turning unstructured content (YouTube) into structured arguments
• enabling comparison, counterpoints, and a personal reasoning trace

What’s next

This site is now my growth log. Instead of hiding work in private repos, I want to:

• publish project write-ups
• keep lightweight demo snapshots
• document decisions and learnings (even the wrong ones)

If you’re reading this during an interview: feel free to click around the Projects and Lab pages.